Fast and non-destructive X-ray fluorescence (XRF) core scanning provides high-resolution element data that are widely used in paleoclimate studies. However, various matrix and specimen effects prevent the use of semi-quantitative raw XRF core-scanning intensities for robust paleoenvironmental interpretations. We present here a case study of a 50.8-m-long piston Core MD12-3432 retrieved from the northern South China Sea. The absorption effect of interstitial water is identified as the major source of deviations between XRF core-scanning intensities and measured element concentrations. The existing two calibration methods, i.e. normalized median-scaled calibration (NMS) and multivariate log-ratio calibration (MLC), are tested with this sequence after the application of water absorption correction. The results indicate that an improvement is still required to appropriately correct the influence of downcore changes in interstitial water content in the long sediment core. Consequently, we implement a new polynomial water content correction in NMS and MLC methods, referred as NPS and P_MLC calibrations. Results calibrated by these two improved methods indicate that the influence of downcore water content changes is now appropriately corrected. We therefore recommend either of the two methods to be applied for robust paleoenvironmental interpretations of major elements measured by XRF-scanning in long sediment sequences with significant downcore interstitial water content changes.